PROJECT SUMMARY (DESCRIPTION) How new mutations affect an individual's fitness lies at the heart of many questions in evolutionary genetics and is highly relevant for elucidating the genetic basis of complex traits. Current evidence suggests that while many mutations are deleterious, selection coefficients for particular mutations come from a distribution of se- lective effects (DSE). Although the DSE has been subjected to intense study over the past decade, many open questions about this distribution and the importance of deleterious mutations in evolution remain unanswered. The widespread use of next-generation sequencing has provided a plethora of new polymorphism datasets that are potentially informative about the nature of fitness effects of mutations. However, existing computational approaches make simplifying assumptions and lack statistical power to extract the full potential of these data. This project will develop and apply new computational approaches to integrate these emerging polymorphism datasets from multiple species together in a rigorous statistical framework to estimate fundamental parameters relating to natural selection across genomes. We will develop novel computational methods that use very large samples of human genomes to better estimates selective effects. First, we propose to leverage runs of homo- zygosity within an individual to co-estimates dominance coefficients and the DSE. We will develop a second set of methods that utilizes transmission patters within large numbers of pedigrees to estimate selective ef- fects. Third, previous studies of the DSE have assumed all sites are independent of each other and have ig- nored epistasis. We will develop a composite likelihood approach to estimate the degree of epistasis between many pairs of SNPs. We will then apply these methods to the large cohorts of human genomes that are being sequenced to obtain the most detailed estimates of the DSE, dominance coefficients, and nature of epistasis from human populations. Next, these new computational tools will be combined with data from a variety of dif- ferent species to address longstanding questions concerning selective effects. We will investigate the evolution of the DSE itself by testing whether it has changed in response to recent shifts in the environment. As an ex- ample, we will test whether the change in selective pressure associated with dog domestication has led to dif- ferences in the DSE between dogs and wolves. This will be accomplished by generating exome sequencing data from of dogs and wolves. Finally, we will investigate why different species appear to show disparate amounts of amino acid adaptation. This work will resolve a current paradox in population genetics and will con- tribute to understanding the different modes of adaptation across different species. In sum, successful comple- tion of this research will provide the most comprehensive estimates of dominance, the DSE, and the amount of epistasis in natural populations to date. These estimates will dramatically improve the interpretation of genetic variation in future evolutionary and medical genetic studies.